JPH0771215B2 - Automatic shooting position determination device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an apparatus for determining a shooting position for shooting an actual object such as a product, a work of art, or an animal or plant from a plurality of directions. The image information photographed from the photographing position determined according to the present invention is accumulated in the center device, and is used for an image information providing device or the like for searching and providing the information to the terminal device via the transmission path.

(Prior Art) When a real object such as a product, a work of art, an animal or a plant is photographed from a plurality of directions, a conventional method for determining the photographing position is to cover an object O as shown in FIG. 7, for example. Assuming a virtual hemisphere OA, there is a method of determining the photographing positions OB at equal angular intervals in the longitude direction λ (horizontal direction) and the latitude direction ψ (longitudinal direction) of this hemisphere.

However, in this case, as the latitude of the virtual hemisphere OA becomes higher (closer to the pole), the density of the shooting position OB becomes higher, and near the pole.
Images taken from OC have virtually the same information (although the orientation of the object in the screen changes). Therefore, it is understood that the distance on the hemisphere between the shooting positions needs to be constant. However, even if a plurality of shooting positions are decided in this way, for images taken from each shooting position,
The difference in the information that a person who sees it substantially changes greatly depending on the shape of the object and the shooting position.

For example, compare an image of a rectangular parallelepiped as shown in FIG. 8 seen from the direction of arrow A (FIG. 9 (a)) with an image seen by rotating it by 10 degrees (FIG. 9 (b)). Compare the difference between the images that you feel when you try and compare the image viewed from the direction of the arrow B (Fig. 10 (a)) with the image viewed by rotating it by 10 degrees (Fig. 10 (b)) The difference when viewed from the direction of the arrow B is clearly larger than the difference when viewed from the direction.

This means that even if the interval between the shooting positions on the hemisphere is constant, the information difference between the images is not constant. Therefore, when these images are displayed one after another in the order in which the photographing positions are adjacent to each other, the information substantially received by the viewer becomes large in one section and becomes small in another section. That is, if such a set of images is stored, the storage efficiency in terms of how much information is given to the viewer becomes poor.

(Object of the Invention) An object of the present invention is to solve the problem that when a plurality of images are taken along a virtual hemisphere covering an object, a difference occurs in information difference between the taken images. An object of the present invention is to provide a device that automatically determines a shooting position where the information difference between images is constant.

(Structure of the Invention) (Differences between Features of the Invention and Conventional Technology) The present invention is directed to the difference in information between the captured images when capturing a plurality of images along a virtual hemisphere covering an object. The most important feature is that the photographing position is determined so that the corresponding surface information change amount is constant between the images. On the other hand, in the conventional technique, the photographing positions are simply determined at equal intervals, which is different from the fact that the amount of information regarding an object between images is not constant.

Hereinafter, the operating principle of the present invention will be described. Here, the above-mentioned surface information change amount is defined as follows. First, as shown in FIG. 3, the shape of an object to be photographed is plane-approximated by a fine polygon ωi (i = 0,1, ... m-1) such as m triangular patches. Let O be a plane-approximated object, and P be a set of points distributed on the surface of the object O with a uniform density. Each element pεP of this point set P is a region r of radius e centered on p
Has (p). However, the radius e is the maximum radius where the regions r (p) do not interfere with each other, and the set of all regions r (p) is R (P).

If the density of the point set P is sufficiently large, the set R (P) of the region r (p) is a set of very fine circular regions distributed on the object O with a uniform density. At this time, each r (p) ∈ R
(P) can be regarded as having the same information amount ε with respect to the object O.

As shown in FIG. 3, when the object O is first photographed from the position u, the projected area of the region r (p) on the screen surface is su.
Suppose Next, when the object O is photographed from the position v, the projected area of the region r (p) on the screen surface is sv. Assuming parallel projection for simplicity, if su
<Sv, the amount of information projected on the screen surface increases by ε ⁺ = (sv-su) ε / πe ² (1) for the region r (p). vice versa
If su> sv, it can be considered that the amount of information decreases by ε ⁻ = (su−sv) ε / πe ² (2). Therefore, the amount of change in the information amount of the area r (p) projected on the screen surface between the two images captured from the positions u and v is proportional to Δε = | su−sv | (3). Therefore, if Δε is obtained for all regions r (p) εR (P), the sum E E = ΣΔε (p) (4) for all r (p) εR (P) is the object projected on the screen surface. It is proportional to the change amount of the information amount of O. The surface of the object O is a fine polygon ωi (i = 0,
1, ... m−1), and a region r (p) on ωi
Are distributed with a uniform density, so that the projected area of ωi on the screen surface when the object O is viewed from the position u is Siu, and the projected area when viewed from the position v is Siv, (5) Duv defined like the formula Is directly proportional to E in equation (4). Therefore, Du in this equation (5)
Let v be referred to as the amount of change in surface information between the image of the object O captured from the position u and the image of the object O captured from the position v.

In the present invention, the amount of change in surface information is proportional to the amount of change in the amount of information between two images, and attention is paid to the point that it can be extracted only from the shape information of the object, and the amount of change in surface information is almost constant between adjacent images. The feature is that the difference in information between the images is made constant by selecting the shooting position so that

(Embodiment) FIG. 1 is a block diagram showing the basic configuration of an embodiment of the present invention, in which 1 is a shape information input unit, 2 is a plane approximation unit, 3 is a surface information change amount detection unit, 4 Is a representative position extraction unit, and 5 is a representative photographing position output.

The specific configuration is shown in FIG. 2, and the plane approximation unit 2 includes an approximation unit 21 and a polygon information storage unit 22. The surface information change amount detection unit 3 includes a latitude direction change amount measurement unit 31, a storage unit 321,
An averaging unit 32 including a calculation unit 322 and a latitude direction change amount measuring unit 33. In addition, the representative position extraction unit 4 includes a storage unit 41.
1. A representative latitude extraction unit 41 including an extraction unit 412 and a storage unit 42
1. The representative longitude extracting unit 42 including the extracting unit 422. Then, as the output of the representative shooting position 5, a representative latitude 51 and a representative longitude 52
Is obtained.

To operate this, the shape information of the object is input to the shape information input unit 1. In the plane approximating unit 2, first, the approximating unit 21 reads the shape information of the object from the shape information input unit 1, and the shape of the object is m fine polygons ωi (i = 0,1 ,. ), The shape information of each ωi and its outward normal vector information are stored in the polygon information storage unit 22. The latitude direction change amount measurement unit 31 of the surface information change amount detection unit 3 is a polygon information storage unit 22.
From this, the shape information of the object that is approximated by a polygon in a plane is read out, and based on this, the longitude of the shooting position for shooting the object is fixed at ψ = 0, and the latitude λ is set to a fine angle Δλ. The amount of change D0 (λ) in the surface information in the case of moving in steps is calculated from the equation (5). However, Δλ = π / (2 · C0) (C0
Is a natural number).

Then move longitude ψ by a minute angle Δψ
Calculate Δψ (λ). However, Δψ = 2π / C1 (C1 is a natural number). Repeat this until longitude ψ = 2π,
Latitude direction surface information change amount Dψ (λ) (ψ = Δψ, 2Δ
ψ, ..., 2π, λ = Δλ, ..., π / 2) is calculated, and the averaging unit 3
The data is stored in the second storage unit 321. The calculation unit 322 reads the surface information variation amount Dψ (λ) in the latitude direction accumulated in the accumulation unit 321, and calculates the average value Da (λ) for the route ψ. Is stored in the storage unit 411 of the representative latitude extraction unit 41.

The extraction unit 412 receives the average value Da (λ) (λ = Δλ, 2 from the storage unit 411).
(Δλ, ..., 2π, ..., π / 2) is read out and the integral of Da (λ) is calculated as shown in Fig. 5 by using the predetermined interval value δ of the surface information change amount. Divide the value evenly.

And x latitudes Λi (i = 0, ..., x−
1) is output as the representative latitude 51, and this is output to the longitude direction change amount measuring unit 33. The longitude direction change amount measuring unit 33 reads out shape information obtained by approximating the object shape in a plane from the polygon information storage unit 22, and then outputs Λi (i = 0, ...) From the extraction unit 412 of the representative latitude extraction unit 41. ., x-1).

Then, in the same manner as the latitude direction change amount measurement unit 31, the longitude direction surface information change amount DΛi (ψ) (i = 0, ..., x−1, ψ = Δ)
ψ, 2Δψ, ..., 2π) is measured and stored in the storage unit 421 of the representative longitude extraction unit 42.

Longitudinal surface information change amount DΛi accumulated in the accumulation unit 421
(Ψ) is read out, and similarly to the representative latitude extraction unit 41, the integrated value of DΛi (ψ) is equally divided at δ intervals as shown in FIG. Then, yi longitudes Ψij (j
= 0, ..., yi−1, i = 0, ..., x−1) is obtained and is output from the extraction unit 422 as the representative longitude 52.

In the present embodiment, the interval value δ of the surface information change amount is described as being preset inside the apparatus, but this interval value δ may be input from the outside.

(Effects of the Invention) If the representative latitude and the representative longitude are determined as described above, the object is photographed from the photographing position on the virtual hemisphere determined by the representative latitude and the representative longitude, and the images are between the adjacent images. The amount of change in surface information becomes almost constant. Therefore, compared with the case where the object is photographed at equal intervals, these images can be displayed more smoothly than continuously displayed. Further, when these images are stored in a database or the like, it is possible to improve the storage efficiency in terms of the amount of information that is substantially given to the viewer, as compared with the case where the images are simply shot at equal intervals.

[Brief description of drawings]

FIG. 1 is a block diagram showing the basic configuration of an embodiment of the present invention, FIG. 2 is a block diagram of the specific configuration of FIG. 1, and FIG. 3 is a definition of the surface information change amount of an object O handled by the present invention. And FIG. 4 are explanatory diagrams for determining the surface information change amount by fixing the longitude of the image capturing position of the object and changing the latitude by minute angles, and FIGS. 5 and 6 are surface information change amounts, respectively. Fig. 7 to 10 are explanatory diagrams for equally dividing the integrated values of latitude and longitude by using the interval value of Fig. 7, and Fig. 7 to Fig. 10 are explanatory diagrams when a conventional object is photographed from a plurality of directions. 8 is a conventional method for determining the
FIG. 10 and FIG. 10 show the difference between the images taken from the A and B directions in FIG. 1 ... Shape information input unit, 2 ... Plane approximation unit, 21 ... Approximation unit, 22 ... Polygonal information storage unit, 3 ... Surface information change amount detection unit, 31 ... Latitude direction change amount measurement unit, 32 …… Averaging section,
321 ... Accumulation unit, 322 ... Calculation unit, 33 ... Longitudinal direction change amount measurement unit, 42 ... Representative position extraction unit, 41 ... Representative latitude extraction unit, 411 ... Accumulation unit, 412 ... Extraction unit, 4 ... Representative longitude extractor, 421 ... Accumulator, 422 ... Extractor, 5 ... Representative shooting position, 51 ... Representative latitude, 52 ... Representative longitude.

Claims (1)

[Claims] 1. A shape information input section for inputting three-dimensional shape information of an object, and a plane approximation section for approximating the three-dimensional shape information input from the shape information input section into a plurality of polygonal planes. A first means for obtaining a surface information change amount obtained by virtually photographing an object subjected to plane approximation by the plane approximation unit from two different viewpoint positions on the surface of a virtual hemisphere covering the object. And a second means for sequentially obtaining, by the first means, the amount of change in surface information between two adjacent points when the photographing position is moved by a minute amount in the longitude direction and the latitude direction on the surface of the virtual hemisphere. Means for determining the actual photographing position in the longitude direction and the latitude direction so that the integrated values of the surface information change amounts between the two adjacent points obtained by the second means become uniform. Shooting position characterized by having means of 3 Auto determining device